Artificial Intelligence

A Philosophical Introduction

Jack Copeland

• Publisher: Wiley-Blackwell
• Year: 2015
• ISBN:
  9781119189855
• Edition: 1
• Topic: Philosophie
• Subtopic: Histoire et théorie de la philosophie

Presupposing no familiarity with the technical concepts of either philosophy or computing, this clear introduction reviews the progress made in AI since the inception of the field in 1956. Copeland goes on to analyze what those working in AI must achieve before they can claim to have built a thinking machine and appraises their prospects of succeeding.

There are clear introductions to connectionism and to the language of thought hypothesis which weave together material from philosophy, artificial intelligence and neuroscience. John Searle's attacks on AI and cognitive science are countered and close attention is given to foundational issues, including the nature of computation, Turing Machines, the Church-Turing Thesis and the difference between classical symbol processing and parallel distributed processing. The book also explores the possibility of machines having free will and consciousness and concludes with a discussion of in what sense the human brain may be a computer.

1
The beginnings of Artificial Intelligence: a historical sketch

To begin at the very beginning …

1.1 The arrival of the computer

The story of the invention of the digital computer is a fascinating one. Folklore has it that the computer originated in the United States, but this is not true. Britain, the USA and Germany developed the computer independently at almost exactly the same time. In terms of who got there first, it is Germany that carries off the cup – or more precisely a lone German, Konrad Zuse. He had the world’s first general-purpose programmable computer up and running by the end of 1941.¹ It was a case of third time lucky: two earlier machines that he built in the unlikely setting of his parents’ living room did not quite work. Although Zuse was first past the post, few were aware of his achievement at the time, and Allied restrictions on electronic engineering in post-war Germany put paid to any further development. Neither Zuse nor his ideas played any significant role in the commercial development of the computer.²

The British story begins at Bletchley Park, Buckinghamshire, a top secret wartime establishment which was devoted to breaking the Wehrmacht’s codes. With a staff of brilliant mathematicians and engineers, it was almost inevitable that Bletchley Park should produce something revolutionary – and when it came it was the Colossus, an electronic³ computer for deciphering coded messages.⁴ The first Colossus was installed and working by December 1943, a full two years after Zuse obscurely made history in Berlin. (Some commentators quibble over whether the Colossus was a true computer.⁵ It was designed for just one specific task, cracking codes, and beyond that could do little or nothing. Enthusiasts at Bletchley had tried to coax the Colossus into doing long multiplication but, tantalizingly, this proved to be minutely beyond the frontier of the machine’s capabilities.⁶ Zuse’s computer, on the other hand, could be set up to perform any desired calculating task (provided, of course, the task was not so large as to exhaust the machine’s storage capacities). Zuse’s was a general-purpose computer, while the Colossus was a special-purpose computer.)

After the war the Bletchley group broke up and the action moved north to Manchester. It was here that F.C. Williams, Tom Kilburn and their team built the Manchester University Mark I general-purpose computer. The first program ran in June 1948.⁷ By April 1949 the small prototype had been expanded into a much larger machine.⁸ On the other side of the Atlantic progress ran just slightly slower. The first comparable American machine (called the BINAC) was tested in August 1949.⁹

A Manchester firm, Ferranti Limited, contracted to build a production version of the Manchester Mark I. These machines were the world’s first commercially manufactured electronic stored-program computers. In all, nine were sold. The first was installed in February 1951, just two months before the appearance of the world’s second commercially-available machine, the American UNIVAC.¹⁰

The Ferranti Mark I has the additional distinction of being the first non-human on the planet to write love letters.

Darling Sweetheart,

You are my avid fellow feeling.

My affection curiously clings

to your passionate wish. My

liking yearns to your heart. You

are my wistful sympathy: my

tender liking.

Yours beautifully,

Manchester University Computer¹¹

The American side of the story begins with a machine known as the ENIAC. (The initials stand for Electronic Numerical Integrator and Computer. It is an exceptionless law that the acronyms used to name computers and computer programs never mean anything interesting.) The ENIAC was built at the University of Pennsylvania by John Mauchly and J. Presper Eckert, and first went into operation in November 1945 (nearly two years after the Colossus).

The ENIAC was a programmer’s nightmare – it had to be rewired by hand for each new task. This was a mammoth operation involving thousands of plugs and switches. It would generally take the operators two days to set the machine up for a fresh job.¹² This primitive monkeying about with cables was all that the ENIAC had to offer by way of a programming facility. The Manchester Mark I was much less user-hostile. The Mark I was the world’s first fully electronic stored-program computer.¹³ Setting it up to perform a new job involved simply feeding in a punched paper tape. The machine would copy the programmer’s instructions from the tape and store them in its own memory. Eckert and Mauchly had realised as early as 1943 that it would be beneficial to be able to store the ENIAC’s operating instructions inside it, but the military wanted the ENIAC operational as soon as possible and so exploration of the stored-program concept had to wait.¹⁴

After the ENIAC, Eckert and Mauchly went on to build the BINAC, which was a stored-program machine. They then built the UNIVAC, the first market place offering of the nascent American computer industry. Thereafter the US quickly came to dominate the commercial production of computers. However, history has not been kind to Eckert and Mauchly. In 1972, a prolonged patents struggle between the Honeywell Corporation and Sperry-Univac ended with the judicial decision that ‘Eckert and Mauchly did not themselves first invent the automatic electronic digital computer, but instead derived that subject matter from one Dr John Vincent Atanasoff’.¹⁵ Atanasoff was an American college professor who very nearly succeeded in building a special-purpose electronic computer during the period 1936 to 1942.¹⁶ Unfortunately he never managed to get his machine properly operational, largely because of malfunctions in a cumbersome piece of equipment for storing information on punched cards. Mauchly paid a visit to Atanasoff’s laboratory in 1941, and in the opinion of the judge it was Atanasoff’s ground-breaking work that had led Mauchly and Eckert to the ENIAC. The judicial ruling notwithstanding, a dispute still rages over the extent to which Atanasoff’s ideas influenced Mauchly and Eckert.

This was not the first time that events had dealt Eckert and Mauchly a bitter blow. The months subsequent to the ENIAC becoming operational ought to have been their time of triumph, but in reality they found themselves upstaged by one of their colleagues – a certain John von Neumann. A gifted mathematician and logician, von Neumann has been described as an ‘unearthly genius’.¹⁷ Von Neumann heard of the ENIAC during a chance meeting on a railway station. He was working on the Manhattan Project at Los Alamos at the time, where he applied his great genius to such sinister problems as the calculation of the exact height at which an atomic bomb must explode in order to inflict maximum destruction. He was quick to see the implications of a machine like the ENIAC (‘shell, bomb, and rocket work … progress in the field of propellants and high explosives … aerodynamic and shock wave problems …’).¹⁸ He offered to act as consultant to the Eckert-Mauchly project, and rapidly established himself as national spokesman on the new computer technology. Von Neumann was a pillar of the scientific establishment and his patronage did wonders for the prestige of the ENIAC project, but as a result of his lectures and publications the computer came to be more closely associated with his own name than with the names of the people who had brought it into the world.¹⁹ Von Neumann had a saying that only a man born in Budapest can enter a revolving door behind you and come out in front.²⁰ He himself, of course, was such a man, and it was Eckert and Mauchly who were left behind.

Von Neumann went on to make huge contributions to computer design. He enunciated the fundamental architectural principles to which subsequent generations of computers have adhered. For this reason, standard modern computers are known generically as von Neumann machines. We shall hear more of von Neumann in later chapters.

The name of John von Neumann is linked with the birth of artificial intelligence in another way. In 1956 the most influential of the early efforts at AI programming ran on the JOHNNIAC, a sophisticated computer designed by him.²¹ The program is known affectionately as the Logic Theorist. It was the work of Allen Newell, Cliff Shaw, and Herbert Simon, three of the great frontiersmen of AI.²²

1.2 The Logic Theorist

Logic is a central preoccupation of AI research. The ability to reason logically is, obviously, a fundamental component of human intelligence; if computers are to achieve the status of artificial intelligences, they too must be given the ability to search logically for a problem’s solution. Newell, Shaw and Simon pioneered the study of how this might be done.

Their initial plan was to write a program that could work out its own proofs for theorems of school geometry. Success eluded them, though, largely because it proved unexpectedly difficult to represent geometrical diagrams in a form that the JOHNNIAC could handle.²³ Undeterred, they turned to the related project of programming a computer to search in a logical way for – incestuously – proofs of the theorems of pure...